Chapter 4. Highly Sensitive and Selective Biosensors Based on Functionalized Organic
4.3. Results and Discussion
4.3.1. Device Fabrication and Characterization
OFET-based sensors with AOCB[6] were prepared with bottom-gate top-contact configuration. The DDFTTF thin film (~15 nm thickness) was thermally evaporated onto n-octadecyltrimethoxysilane (OTS)-treated SiO2/Si substrates at an optimal substrate temperature of 105 qC. Source and drain electrodes (~40 nm thickness) were formed by evaporating gold through a shadow mask. In addition, the source and drain electrodes in the channel area were covered with a SiO passivation layer (~20 nm thickness). The SiO layer acts as an electrical insulator and chemical barrier to prevent the source-drain electrodes from peeling off during the OFET sensor operation in liquid solutions.[23, 38, 55] An AOCB[6]
solution (~5 mg mL-1) in methanol was spin-coated to form a stable and homogeneous receptor layer on the semiconductor film for selective analyte adsorption. Further details on the fabrication of OFET- based sensors are described in the Experimental Section. The corresponding device structure and AOCB[6] are shown in Figure 4.1a and 4.1b, respectively. The electrical characteristics of OFETs with and without AOCB[6] were measured in the saturation regime as shown in Figure 4.2. The DDFTTF OFETs without AOCB[6] had an average field-effect mobility (PFET) of 0.053 cm2 V1 s1, with an on/off current ratio (Ion/Ioff) of more than 106. After functionalization with the AOCB[6] layer, the DDFTTF OFETs showed an average PFET of 0.028 cm2 V1 s1 and Ion/Ioff larger than 105. The electrical performances of DDFTTF OFETs fabricated with and without AOCB[6] are summarized in Table 4.1.
The mobility degradation after functionalization with AOCB[6] may arise from the electron-donating effect of AOCB[6] molecules that decreases the density of holes (charge carriers of p-channel devices), as well as from the trapped impurities generated during solution processing.[56] The off-current increased by one order of magnitude most likely due to the effect of doping by oxygen, leading to the decreased Ion/Ioff under ambient conditions.[57] The threshold voltage (VTH) was changed from 26.3 V to 5.6 V, indicative of the easier turn-on after functionalization with AOCB[6]. Despite the minor degradation in the charge carrier mobility of the OFET devices, they showed ample device performance for sensing analytes in the aqueous phase (vide infra).
4.3.2. Thin-Film Microstructure Analysis
69
We investigated the morphological characteristics of organic thin films using atomic force microscopy (AFM) analysis (see Figure 4.3). The DDFTTF thin films showed various grain sizes and distinct grain boundaries with a relatively large surface roughness (a root-mean-square (RMS) roughness of 7.5 nm) (Figure 4.3a and 4.3b). The thin films became smoother with a RMS roughness of 1.8 nm after thermal annealing at 150 qC (Figure 4.3c and 4.3d). The AOCB[6] layer spin-coated on DDFTTF film was annealed at 150 °C in a nitrogen atmosphere to remove the residual solvent. The thermogravimetric analysis (TGA) of AOCB[6] revealed that it was stable up to ~310 qC. The AOCB[6]
layer covered the DDFTTF device uniformly and completely with a RMS roughness of 2.5 nm (Figure 4.3e and 4.3f), and the cross-sectional AFM analysis revealed that the thickness of the AOCB[6] layer was 17.0 nm. As the height and diameter of AOCB[6] was about 0.9 nm and 2.2 nm, respectively, it is considered that approximately 10 or more layers of AOCB[6] were deposited with a high density on the semiconductor film.
4.3.3. Sensitivity and Selectivity of OFET-Based Sensors
In OFET-based sensors, chemical or physical adsorption of target analytes leads to a change in the channel current, which depends on the analyte composition, concentration, and OFET operating conditions.[24, 58-60] In addition, the OFET-based sensors have excellent current-amplifying properties induced by an external gate field. A sensing platform was prepared by placing a polydimethylsiloxane (PDMS) mold reservoir onto the OFET sensor device, and sensing experiments were performed under ambient conditions. Prior to detecting the analytes, a baseline current was estimated with deionized (DI) water. The DDFTTF OFET-based sensors exhibited minimal sensing signals upon continuous exposure to DI water, as shown in Figure 4.4a. After stabilizing the drain current, solutions (~15 μL) containing analytes were injected into the PDMS reservoir. The sensitivity of sensors was calculated by dividing the measured data by the baseline current.
Figure 4.4a shows the liquid-phase sensing behaviors of DDFTTF OFET-based sensors functionalized with AOCB[6] toward ACh+ and Ch+. The sensors showed positive sensing behaviors, in which the drain current was enhanced after injection of the analytes. Surprisingly, the detection limit (1 pM) of the DDFTTF sensors with AOCB[6] toward ACh+ was six orders of magnitude lower than that (~μM) of ISE-based sensors[36] and two orders of magnitude lower than that (100 pM) of AChE- based biosensors,[35] respectively. We also monitored changes in the drain current of the sensors with and without the AOCB[6] layer, while the devices were exposed to 1 pM of ACh+ (Figure 4.5). The sensors with AOCB[6] showed much higher sensitivity for ACh+ compared to the sensors without AOCB[6]. The OFET-based sensors without AOCB[6] exhibited low sensitivity because the grain boundary defects in organic semiconductors could solely provide pathways for the diffusion of analytes into the channel region. These results indicate that introduction of the AOCB[6] layer significantly improves the sensitivity of the sensors through selective binding of ACh+ on the device surface.
In addition, responses of the sensors with AOCB[6] toward sodium ion (Na+), which is also known 70
as interfering species for the detection of ACh+, were monitored (Figure 4.6a). Interestingly, the sensing signals of Ch+ and Na+ were very different from ACh+. For Ch+, almost no signals were detected at concentrations lower than μM. Moreover, no signals were detected for Na+ at μM concentrations.
Figure 4.4b shows the statistical comparison of the sensing data for the concentration of ACh+ and Ch+. The change in the drain current for ACh+ was observed at a wide concentration range (1 pM – 100 mM), whereas the sensors for Ch+ and Na+ (Figure 4.6b) exhibited no detectable signals at concentrations lower than 1 μM. These results support the excellence of our sensor devices for the selective and sensitive detection of ACh+. Such superior sensing ability of our sensor devices originates from the commendable combination of highyl selective synthetic receptors and highly sensitive OFET devices.
We also monitored sensing signals of ACh+ by using a baseline buffer solution, instead of DI water (Figure 4.7). For the sensing test, acetylcholine chloride solutions were prepared in a phosphate- buffered saline (PBS) solution (pH 7.4, 0.01 M) containing sodium chloride (137 mM) and potassium chloride (2.7 mM). This sensing condition was close to the normal physiological conditions found in blood. The AOCB[6]-functionalized OFET sensors could also detect ACh+ with the detection limit down to 1 pM in the PBS solution, although they exhibited relatively lower sensitivity compared with that in DI water due to the interfering effects of cations.
To understand the superior selective sensing nature of our sensor devices, we performed density functional theory (DFT) calculations of AOCB[6] with these three analytes. In these calculations, CB[6]
was used as a host material because the guest binding nature of AOCB[6] is essentially the same as CB[6]. Note that the allyloxy group is only introduced for solvent orthogonality. DFT optimized structures of ACh+ and Ch+ interacting with the CB[6] host are shown in Figure 4.8. In the energy- minimized configuration, ACh+ is parallel to the cavity axis and the acetyloxy end group was penetrated inside the cavity with its methyl group interacting with one of the carbonyl portals. The positively charged trimethylammonium group was located over the opposite carbonyl portal. Thus, complexation between CB[6] and ACh+ showed a strong binding energy of 86.5 kcal mol1, which is attributed to the strong charge-diople interactions and hydrogen bonding between CB[6] and ACh+. When AOCB[6]
forms a complex with ACh+, the carbonyl group of AOCB[6] would partially donate electrons to the positively charged ammonium group of ACh+, and these charge-dipole interactions tend to increase the electron-withdrawing characteristics into the channel region, thereby leading to an increase in the hole current of the p-channel OFET sensor devices. For complexation between CB[6] and Ch+, the ammonium group of Ch+ weakly interacts with the portal of the host and the hydroxyethyl group of Ch+ does not enter the host cavity. Thus, AOCB[6] forms a weak host-guest complex with Ch+, which showed a lower binding energy of 69.5 kcal mol1 than that of CB[6]-ACh+. The calculated structures were well matched with the 1H-NMR spectra of each complex.[36, 37] In 1H-NMR, protons of trimethylammonium groups of ACh+ show small downfield shifts and protons of the acetyl group show large upfield shifts upon complexation with AOCB[6], which indicate the formation of strong host-
71
guest complexes as shown in Figure 4.8a. However, all protons of Ch+ show very small downfield shifts upon addition of AOCB[6], which indicates that the ammonium group of Ch+ interacts weakly with the portals of the host and the hydroxyethyl of the molecule exists outside the cavity, as shown in Figure 4.8b. Forcomplexation between CB[6] and Na+, Na+ interacts weakly at the portal of the host (Figure 4.9).
Since DDFTTF OFET-based sensors are p-channel devices, complexation of the host molecule with cationic guest molecules would increase the signals by increasing the hole carrier density in the channel region owing to the overall electron-withdrawing characteristics toward the active layer. Because all three analytes have the same +1 charge, we assume that the differences between sensing behaviors of guest molecules should be related to their binding constants to the host molecule. In addition to this, the partial charge change differences of the host molecule upon complexation with each guest molecule would also affect the differences in the sensing signals. Therefore, we calculated the partial charge changes in the host molecule upon complexation with each guest molecule (Figure 4.10). As in the case of sensing signals, complexation with ACh+ shows the largest charge changes on CB[6] (0.063 r0.008), Na+ shows the smallest charge changes (0.020 r0.005), and Ch+ shows moderate changes (0.052 r0.009). Considering the sensing results of each molecule, the difference in charge changes between ACh+ and Ch+ should be larger than the value shown in Figure 4.10. The reason for the difference between calculated and experimental results may originate from the differences in binding probabilities (i.e. binding constant) of each guest molecule to the host molecule on the device surface. In contrast to ACh+, which strongly binds to the host, Ch+ has a very weak interaction with the host. These large differences in the binding probability and accumulation of each binding event on the device surface may cause such a large difference in the sensing signals.
For ACh+ sensors, a significant analytical challenge is the detection of ACh+ with high sensitivity and selectivity in the presence of Ch+. Thus, the signal intensities of the sensors for analyte blend systems were investigated using a mixture solution of two analytes (Figure 4.11). Mixed solutions containing 1 μM ACh+ and Ch+ analytes were prepared with various volume ratios (1:1, 2:1, and 3:1) of ACh+ relative to Ch+. The sensors with AOCB[6] showed enhanced signal intensities with an increasing volume ratio of ACh+ in the mixed solutions, indicative of the high selectivity toward ACh+ compared to Ch+. This was due to the relatively higher binding affinity of AOCB[6] toward ACh+ compared to Ch+.
In addition to amplification of the detected signals, OFET-based sensors are suitable for applications in low-cost and flexible electronics. To explore the possibility of using a flexible sensor platform, our sensors were also fabricated with indium tin oxide (ITO)-coated polyethylene naphthalate (PEN) as the polymer substrate and aluminum oxide (Al2O3) as the transparent dielectric (Figure 4.12a). A 100 nm- thick Al2O3 gate dielectric layer was deposited on the PET substrate via a radio frequency (RF) magnetron sputtering technique, and a photograph of the resulting flexible sensor is shown in Figure
72
4.12b. The transfer and output characteristics of the DDFTTF OFET-based sensor with AOCB[6] are shown in Figure 4.13a and 4.13b, respectively. The results of the sensing experiments for ACh+ exhibited performances similar to SiO2 dielectric-based sensors. The flexible sensors could also detect ACh+ with a detection limit of 1 pM under low-voltage operation conditions (VDS = 0.5 V and VGS = 10 V) (Figure 4.12c). These results describe the first demonstration of ACh+ sensing without any enzymatic reactions using synthetic receptor-functionalized flexible FET-type sensors. In addition, our findings expand the range of practical applications of OFET-based sensors.